88Z AI, 3 May — xAI has released Grok 4.3, positioned as a pragmatic transitional version. The model features significantly reduced API pricing, faster output speeds, and support for million-token long contexts, with enhanced performance in document generation and office assistance tasks. While showing marked improvement in agent tasks, it still lags behind top competitors in complex reasoning and fact-checking stability, with an increased hallucination rate.
At a glance
- Grok 4.3 cuts prices, boosts speed, supports million-token context, and enhances office assistance
- Complex reasoning and fact-checking stability lag behind top competitors, with higher hallucination rates
The update
xAI has officially launched its latest flagship reasoning model, Grok 4.3, defining it as a pragmatic transitional product. In terms of cost and efficiency, the model's API prices have been reduced by 40% to 60%, with output speeds reaching 196 tokens per second and support for context windows of up to one million tokens. Functionally, Grok 4.3 shows significant improvements in agent tasks and office assistance, excelling particularly in generating documents, tables, and presentations. However, in intelligence index rankings, the model still trails behind GPT-5.5 and Claude Opus 4.7. It exhibits deficiencies in complex reasoning capabilities and the stability of fact-checking, while also showing an increased hallucination rate, indicating room for improvement in high-end reasoning scenarios.
Sources & editorial note
This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.
- 新智元Source published
Source dates above refer to the original announcements; the date under the headline is this article’s publication on 88Z AI.